Analytica

Pros: Influence diagrams often capture the essence of models. They can be readily communicated or shared. The insights are often more valid than any particular computational results.

Even complex systems can often be represented comprehensibly with array mathematics. Analytica's declarative programming and array mathematics make it possible for the program to be structurally very similar to the underlying conceptual model. This improves the ability to review and test models, to communicate them, to maintain them, and to share and re-use them.

The same features mean that a model can be used from the outset for broad n-dimensional parametric uncertainty analysis ("exploratory analysis under deep uncertainty), which would otherwise require a good deal of programming that tends to be deferred and--often--never finished. Such uncertainty analysis is a core element of modern policy analysis for systems in which many things are deeply uncertain and so-called best-estimate point calculations are of dubious value. Analytica also has excellent built-in features for probabilistic modeling and Monte Carlo calculations.

Analytica can be used for many types of modeling, including time-stepped simulation and MIT-style System Dynamics. It can be used for logic models and qualitative models.

Cons: Analytica is not an "agent-based language." Nor is it suitable for discrete-event simulation. Its less expensive Its lower-cost "Pro" version does not incorporate optimization such as Excel's Solver (the advanced versions of Analytica do). Although it supports statistical analysis, Analytica is not polished for such work. Nor is its graphics package as powerful as specialized systems such as Tableau.

Currently, Analytica does not play easily with modules written in Python, R, etc. Also, only the more expensive versions of Analytica allow Analytica models to be used as submodes within much bigger models that might include, e.g., agents, system dynamics, and discrete-event simulation.

Analytica does not have built-in features for configuration control, validation and verification, etc.

Overall: Analytica is very powerful. It's not for everyone, but should be viewed seriously for those interested in comprehensibility, uncertainty analysis, and other elements of analysis.

It is well suited for quick learning and relatively simple modeling, or for much more complicated modeling, but the latter has some steep learning curves and is not for the tyro. Professional programmers may prefer writing in their personal favorite lower-level languages.